Читать книгу Service Level Management in Emerging Environments - Nader Mbarek - Страница 47

Various projects and research studies have been carried out with the aim of studying the guarantee of QoS in the IoT environment based on a transversal approach involving the different layers of the IoT architecture. We present below a non-exhaustive list of proposals produced by these studies.

SymbIoTe, a European project funded by H2020 (January 2016–January 2019) (Zarko et al. 2015), aims to offer a system providing IoT services if the availability and QoS potentially associated with this can be guaranteed by the underlying infrastructure. The IoT system must prioritize privileged users over basic users if there are simultaneous attempts to access certain IoT services. This priority must be ensured across all IoT layers in order to guarantee end-to-end QoS. In addition, the system must allow access control to IoT services in accordance with local and global legislation (security of access to critical data is a necessary requirement requested by several governments and international organizations), the current overload, security issues, etc. This access control must be provided across all IoT layers (Zarko et al. 2015).

In addition, several research projects have studied the transversal approach to guaranteeing QoS in the IoT environment. The research work presented in Duan et al. (2011), for example, is based on the principle that two adjacent layers in the IoT architecture communicate with each other to provide QoS. The upper layers send a QoS query to the layer that is directly beneath, which translates the request into a number of QoS parameters. This interlayer communication is carried out using dedicated interfaces with clearly specified parameters. The QoS for the support and application layer in the IoT architecture is directly perceived by the clients and is based on parameters such as time of service, precision, load and priority. The network layer QoS depends on the network type and is based on parameters like bandwidth, delay, jitter, and packet loss ratio, etc. Finally, QoS for the device layer is based on parameters of sampling, coverage, synchronization and mobility. It is useful to differentiate services in order to identify different service levels and attribute these to clearly defined IoT applications. In addition, each level corresponds to a number of specific QoS requirements. The research work conducted in Duan et al. (2011) organized applications into four classes depending on the type of task to be performed. The four categories of applications specified are as follows: Guaranteed Service, Guaranteed Service/Differentiated Service, Differentiated Service, and Best Effort. Therefore, IoT control applications require Guaranteed Service, IoT queries require Guaranteed and Differentiated services, IoT real-time surveillance applications require Differentiated Services, while non-realtime surveillance applications require Best Effort services.